Hard computer Crash AMD

[Gordhead]

Hello I've been having a strange problem for quite some time now (about 2 years). I have an AM4 platform that is hard crashing during gaming. It never shuts off during regular tasks. Computer completely shuts off and the RGB stays on but the computer is off. The only way to reset is to hard reset the computer. I've tried multiple different solutions, ram, gpu, power supply, bios updating, reinstalled OS, DDU, reseating CPU, reseating m.2 drives, I even took out my commander pro just to see if I was overvolting the 12v rail. Nothing seems to be working. What's really odd is how strange and erratic the behavior is, sometimes I can go months without it doing it and then suddenly every time i launch a game to play, 5 mins in it will crash. Let the computer sit off and clear the memory and then no more problems for a bit. I have provided screenshots of benchmarks that will run for hours with no thermal issues or power issues. The only "overclock" I have is on the GPU memory which the 6900xts are known to take very well. I took the overclock off for troubleshooting and it would still do it even at default settings. No overclock on CPU, XMP is enabled on ram. I play a variety of games, Destiny 2, league of legends, American truck sim, dead cells, along those lines, usually nothing too serious. The 2d games seem to run great and never crash, 3d games seem to crash a bit (this is why i figured it was GPU), swapped in my old radeon VII, and my old rx 480 and still got crashes. Oddly enough I played through the entirety of cyberpunk with 0 crashes.  I get no error messages in windows error log, nothing in corsairs gui, nothing in amd. The only thing I get is Bios/amd gpu have been reset to default when I reboot. The only other thing I can think of is the CMOS battery is going out but it retains my settings upon shutdown. Attached some screenshots with temps and power for both CPU, GPU, both were fully soaked when I took the screenshots, benchmarks ran for about an hour a piece. I posted about this a while back and thought my smart access memory was causing issues but it wasn't.  Any help would be greatly appreciated!

 

My System

CPU-5900X

Motherboard- AORUS ELITE X570

AIO-CORSAIR H150I 360 AIO

GPU- ASUS TUF 6900XT

RAM- 32GB KIT 3600 CORSAIR VENGENCE- AMD OPTIMIEZED

POWER SUPPLY- RM1000I (1000W)

STORAGES- SAMSUNG 980 1TB (OS), SAMSUNG 980 PRO 2TB (GAME), 4TB SEAGATE BARRCUDA HARD DRIVE (BULK STORAGE)

PCIE CARDS- TPLINK WIFI 6. 

ACCESSORIES- CORSAIR COMMANDER PRO (OLD ONE), 6 LL CORSAIR FANS, COUPLE CORSAIR RGB STRIPS.

 

Windows 11, 64 bit, up to date

Bios F39

 

4K MONITOR, 1440P SIDE MONITOR, BOTH USING DISPLAYPORT.

[PDifolco]

Did you check GPU temps in gaming ?
Your temps while running Unigine are fine but some games have more or less effect on GPU temps, plus game sessions are longer...

Adrenaline setting reset tends to show a graphic driver failure, but your BIOS shouldn't be reset unless it's a RAM issue and BIOS can't even post, so I'm a bit confused 

Also I had weird crashes in early versions of BG3 (and it was every 3 minutes!!), and in some Unity games,  that went away as soon as I limited my max GPU clock to 2400MHz (on a 7900XTX boosting up to 2800/2900), you could try something like this

[Gordhead]

Hello, yes its about 5-10c hotter or lower depending on the game. Max temp is usually 85C on hotspot. That's why I thought it was GPU at first by my other GPUS did the same thing that's why it was strange. When I reboot I'd go into bios and get the your bios has been reset to defaults screen. Once in windows I'd open amd adrenaline and i would get the watt man has been reset to defaults message. I'll set the max freq and see if that helps. BG3 did the same thing to me as well. 

[ItTakes2ToMango]

85C is fine, shouldn't crash until at least 95C+ if it's purely a thermal problem. 

 

Is that the hottest temp on the GPU? Any other hot spots on the vram or something?

[Gordhead]

Yes That's the hottest I've ever seen it get was 85C (overall) on GPU.

 

CPU never gets above 65C while gaming. VRAM is good on GPU, vrm is good on motherboard as well. I'd have to run another test to get exact numbers. 

[PDifolco]

44 minutes ago, Gordhead said:

Hello, yes its about 5-10c hotter or lower depending on the game. Max temp is usually 85C on hotspot. That's why I thought it was GPU at first by my other GPUS did the same thing that's why it was strange. When I reboot I'd go into bios and get the your bios has been reset to defaults screen. Once in windows I'd open amd adrenaline and i would get the watt man has been reset to defaults message. I'll set the max freq and see if that helps. BG3 did the same thing to me as well. 

Then I would rather think of a RAM issue, GPU crashes don't reset BIOS afaik

Can you disable XMP and test if it still crashes ? AMD weakness is RAM controller, and I have had huge issues with Corsair Vengeance RAM on my old DDR4 rig, same 5900X, same 2x16GB sticks, and nearlythe same board than yours (Aorus X570 Pro ), had to RMA a kit

[nether1234]

I have had my computer hard reset a few times since I got my 7800XT last year. Every time it did it I noticed the GPU driver needed an update. Now I make sure it's always updated before I game and haven't had an issue since.

 

Have you also checked your PSU to make sure it's still good? Do you have a PSU tester or spare PSU you could test with the system?

[Gordhead]

Its very possible its ram. But I tried my other kit of gskill triden 3200mhz. and it still did it. My other psu did it as well. I may just have to deal with it unfortunately.  It may just be the ram controller on my board acting up and may need to manually go to 3600 instead of using xmp profile. Ill have to test to see if its resolved. Like I said it can disappear for months than spring back up non stop. Currently its not doing it. Its very weird. 

[PowerPCFan]

3 hours ago, Gordhead said:

Hello, yes its about 5-10c hotter or lower depending on the game. Max temp is usually 85C on hotspot. That's why I thought it was GPU at first by my other GPUS did the same thing that's why it was strange. When I reboot I'd go into bios and get the your bios has been reset to defaults screen. Once in windows I'd open amd adrenaline and i would get the watt man has been reset to defaults message. I'll set the max freq and see if that helps. BG3 did the same thing to me as well. 

85c hotspot is fine, my 6700 XT regularly hits 90c hotspot with the default fan curve in the hot summer. 

[SpookyCitrus]

With everything you have troubleshooted and tried, I'd say your likely looking at a bad CPU or motherboard. As those seem to be the only two parts you haven't entirely replaced for testing.

 

Another possibility I thought of while typing this is a failing SSD. A bad storage drive could cause hard crashes as well, have you tested the smart status of your drives? When you reinstalled Windows did you put it back onto the same Nvme or have you tried completely different drives as well?

[Gordhead]

Sorry I've been absent from my issue, think I finally figured it out. My CMOS battery seemed malfunctioning and that's why my bios was resetting. I wasn't paying attention to this but also my time was resetting in the bios, windows time was fine since it synced with internet. After doing a bunch of research apparently there was a bad batch of cmos battery's making their way through the Aorus elite boards. Since changing my battery I haven't come across any more issues with bios resetting/computer crashing.